CO2 Displacements of Reservoir Oils From Long Berea Cores: Laboratory and Simulation Results
- V.J. Kremesec Jr. (Amoco Production Co.) | H.M. Sebastian (Amoco Production Co.)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- May 1988
- Document Type
- Journal Paper
- 496 - 504
- 1988. Society of Petroleum Engineers
- 1.6.9 Coring, Fishing, 4.3.4 Scale, 5.5 Reservoir Simulation, 5.4.1 Waterflooding, 5.2.1 Phase Behavior and PVT Measurements, 5.2 Reservoir Fluid Dynamics, 4.1.5 Processing Equipment, 4.1.2 Separation and Treating, 5.2.2 Fluid Modeling, Equations of State
- 2 in the last 30 days
- 379 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
Summary. CO2 displacements of three different reservoir oils from long Berea cores were conducted over a wide range of reservoir conditions with the pressure always above the slim-tube minimum miscibility pressure (MMP). The detailed performance of these displacements is simulated with a fully compositional simulator and the Redlich-Kwong (RK) equation of state (EOS). Oil recovery, GOR, and effluent profiles are compared with experimental results. The EOS is observed to be capable of predicting the phase-behavior transitions that occur in situ when miscibility is generated phase-behavior transitions that occur in situ when miscibility is generated by multiple contacts. The good comparison between experimental results and the simulation has led to specification of a minimum data set for which an EOS should be able to predict before a priori simulations of displacement experiments can be made. The simulation assumed a stable one-dimensional (1D) displacement and that phase-behavior effects play the primary role in the displacement, but the number of gridblocks had to be adjusted in a rough correlation with the CO2/oil-viscosity ratio. Experimentally, final oil saturation and CO2 breakthrough time also correlate with the viscosity ratio. This suggests that viscous instabilities play a role in the experimental displacement, but in this geometry and under these displacement conditions, they can be simulated as numerical dispersion.
CO2 flooding of oil reservoirs is now being conducted on a commercial scale in west Texas and is being implemented in the Rocky Mountains. These projects are based on many factors, including oil recovery predictions. This paper discusses the ability to predict laboratory core displacements. It includes as many of the laboratory measurements and observations as possible: phase-equilibrium measurements, slim-tube MMP and oil recovery, GOR, core effluent properties, and sight-glass observations for displacement tests in Berea cores. An ability to predict or match this comprehensive list of data has not been shown previously. In addition, building such a capability also improves the mechanistic understanding of the process and suggests additional simulator model requirements and experimental data to be collected.
Oil recovery by CO2 depends primarily on the CO2/oil phase behavior, dispersion, and multiphase fluid flow, including the effects of instabilities that arise because the fluids have different den-sities and viscosities. Above the MMP, CO2 displacements oreservoir oils generate miscibility by multiple contacts. The gener ation of multiple-contact miscibility (MCM) means that the fluid compositions, densities, and viscosities are continually changing in situ and there are transitions in phase behavior. These fluid properties and phase-behavior transitions determine the effectiveness properties and phase-behavior transitions determine the effectiveness of the displacement in the laboratory or the field. To simulate these phase-behavior transitions, a fully compositional simulator with an phase-behavior transitions, a fully compositional simulator with an EOS is necessary.
For slim-tube displacements, fluid flow is stable and dispersion is very low; therefore, oil displacement should depend mainly on phase behavior. This dependency has been discussed in several phase behavior. This dependency has been discussed in several papers that have shown good predictions of slim-tube oil papers that have shown good predictions of slim-tube oil recovery. Gardner et al. and Orr et al. used ternary representations of the equilibrium phase-behavior measurements of the CO2/reservoir-oil systems and empirical mixing rules for density and viscosity. Sigmund et al. and Nghiem and Li used a fully compositional simulator with the Peng-Robinson (PR) EOS adjusted to match equilibrium phase behavior. Ngheim and Li computed densities with the PR EOS and viscosities with a Stiel-Thodos correlation. Sigmund et al. computed densities with the PR EOS; viscosity calculations were not discussed.
Core tests have two advantages over slim tubes, even though it is recognized that dispersion and instabilities may affect performance. Core tests can be more representative of the reservoir performance. Core tests can be more representative of the reservoir process, and they provide an opportunity to collect effluent samples from process, and they provide an opportunity to collect effluent samples from which compositions, densities, and viscosities can be determined. Good prediction of these effluent properties is desirable because it gives confidence that the predicted in-situ properties are correct. Good prediction of ultimate oil recovery from any medium is not a satisfactory measure of the ability to predict in-situ properties.
Two attempts have been made to simulate CO2 displacements of oil recovery in laboratory corefloods. Leach and Yellig simulated the displacement of a synthetic oil by CO2 from an 8-ft [2.4-m] Berea core. They used a fully compositional simulator and a version of the RK EOS. The number of gridblocks was adjusted so that the slope and shape of the effluent compositional profiles were matched. They successfully simulated oil recovery and effluent compositions. In retrospect, one can conclude that their dis-placement was dominated by phase-behavior effects with fluid-flow effects playing a secondary role. This paper follows in the foot-steps of Leach and Yellig, but addresses reservoir oil displacements and uses an improved version of the EOS and simulator. In contrast, Gardner and Ypma simulated core displacements that are dominated by viscous fingering. Phase behavior was described with a ternary diagram. This is an important paper, but it deals with experiments and fluid-flow models that are quite a bit removed from the present study.
In this paper, an ability to simulate the main features--oil recovery, GOR, and the component separation and general shape of the effluent profiles--of CO2 displacements of three different reservoir oils from long Berea cores is demonstrated. The three reservoir oils represent a wide range of pressure, temperature, composition, and physical properties. All core displacement experiments were conducted at reservoir temperature, average reservoir rates, and pressures above the slim-tube MMP. The simulations were made with a generalized compositional model described by Young and Stephenson with a version of the RK EOS presented by Turek et al., whose parameters had been tuned to match equilibrium phase-behavior data and slim-tube MMP. The simulator was operated in a 1D mode, and the assumptions that the displacement is stable and thermodynamic equilibrium exists throughout the core were inherent in it.
For all three oils, a good comparison between predicted and experimental oil recovery and GOR was obtained. The predicted effluent compositions are in good agreement with the general shape and separation of components.
|File Size||926 KB||Number of Pages||9|